1. Field of the Invention
The invention relates to a method and device for encoding and reconstructing computer-generated large-area video holograms using a display with conventional (i.e. commercially available) resolution; the display gives a large viewing angle and a high spatial image quality. The display includes a holographic array with controllable pixels, which electronically affect the amplitude and/or phase of light. Such an array is known as spatial light modulator (SLM). A suitable array for spatial amplitude modulation of a light pattern to reconstruct video holograms is, for example a Liquid Crystal Display (LCD). However, this invention can also be applied to other controllable holographic arrays, which use coherent light for modulating a light wave front.
2. Definition of Terms
The term “pitch” describes in this document the distance between the centers of two adjacent pixels of an array. It thus characterizes the display resolution.
The term “encoding” describes the way in which the holographic array is supplied with control values so that it reconstructs a three-dimensional scene. According to this invention, the scene is viewable through ‘viewing windows’.
A viewing window is the intersection of a viewing zone and an observer plane. The observer can see the reconstructed object if at least one eye is inside the viewing window.
3. Description of Related Art
A drawback of 3D-autostereoscopic displays using conventional optics is a mismatch between parallax information and the accommodation of the lens of the eye. On the one hand, each eye sees a different perspective view of a scene, which simulates a depth impression of objects at an arbitrary distance. On the other hand, each perspective view is located on the display surface itself Hence, the eye focuses on the display surface, and each eye sees a flat image. This causes a mismatch between seeing objects at arbitrary depth (i.e. not on the display surface) achieved by parallax information and the accommodation of the eyes to the fixed display surface. This may cause an unpleasant feeling and eye fatigue. This can be avoided by using holographic displays, which reconstruct the objects of a 3D scene at correct depths.
A holographic display reconstructs objects by coherent superposition of light waves. For this purpose, a spatial light modulator (SLM) generates a wave pattern. This hologram is the Fresnel transform of the 3D scene which is to be reconstructed. The SLM diffracts the light waves and reconstructs the scene. As the hologram is sampled, a periodic replication occurs that is associated with a periodicity interval. Thus, the observer within a viewing zone given by the periodicity interval can see the reconstruction. The maximum diffraction angle of the SLM, which basically depends on the pixel pitch, determines the viewing zone. A major problem in encoding and reconstructing video holograms is that a sufficiently large viewing zone must be provided for viewing the reconstruction.
In conventional holographic displays, the viewing zone should cover at least the eye separation, which requires a pixel size of ca. 10 μm at most. Even for a small display area of 100 mm*100 mm, the pixel count will be of the order of 100 million. This causes costly hardware and long computation times even when the display is simplified to a horizontal-parallax only hologram. Currently available large-area displays typically use holographic arrays with a pixel pitch which only diffracts light into a very small viewing zone, so that it is impossible to view a reconstructed three-dimensional scene with both eyes.
Several solutions are known to solve these problems.
The document, K. Maeno, N. Fukaya, O. Nishikawal, “Electro-holographic display using 15 Mega pixels LCD”, Advanced 3D Telecommunication Project, 1996, SPIE, Vol. 2652, refers to holographic 3D displays using commercially available Liquid Crystal Displays (LCD). This document describes a video hologram-reconstructing device that uses five special displays with a high resolution (instead of a conventional LCD with a low resolution); the viewing zone is enlarged because the resolution values of each display is combined to give an overall high resolution. All displays are joined either directly or by way of optical reproduction. Only horizontal parallax is used; vertical parallax is disregarded. The known solution requires a resolution of 15 Mega pixels provided in a unit of the five special displays: each has 3,200×960 pixels to reconstruct a video hologram in a volume of only 50 mm×150 mm and a depth of 50 mm. The viewing zone is only 65 mm wide, corresponding to about the eye to eye separation distance, so that the scene can only just be viewed with both eyes. The required resolution depends on the desired size of the video hologram and the viewing zone. However, this arrangement has significant disadvantages: use of multiple displays and a large lens for reconstruction, leading to a large depth and large volume, and heavy demands on computing power.
Another way of enlarging the viewing zone is described in the document by T. Mishina, M Okui, F. Okano, “Viewing zone enlargement method for sampled hologram that uses high-order diffraction”, Applied Optics, 2002, Vol. 41, No. 8. According to this method, not only the first diffraction order is used for hologram reconstruction, but also further diffraction orders; these are combined to form a common viewing space. The corresponding video holograms for a certain object are shown sequentially on a LC display. With the help of a second LC display, which acts as a spatial frequency filter, the individual diffraction orders are filtered during reconstruction. The visible areas are generated sequentially and joined spatially. The achievable viewing zone is still narrower than 65 mm, so that the reconstructed object can only be viewed with one eye. Again, this method has the disadvantage of heavy demands on computing power. In addition, the pixel arrays are required to have extremely short switching delays.
When joining several diffraction orders sequentially, the displays used must have a high resolution and a high switching speed to prevent the image from flickering. This is why binary holograms are often used. However, they suffer from substantial errors caused by binary encoding.
A common additional disadvantage of the cited known holographic methods is the heavy demand on computing power for encoding the holograms.
A device described in Document WO 2003/021363 (A1) for reconstructing computer-generated holograms contains a vertically oriented line light source that generates holograms with horizontal parallax only. The line light source generates monochromatic light with a bandwidth of less than 10 nm and which is coherent in the horizontal direction but incoherent in the vertical direction.
In conventional holographic displays, the viewing window is much larger than the pupil of the eye is. A consequence is much effort is done to project light into regions where no observer is located.
A basic idea of applicant's former patent application WO 2004/044659 (A2) is to reduce the viewing window to a size that is just slightly larger than the pupil of an eye. This will significantly lessen the requirements on the maximum pixel size. The document describes a device for reconstructing video holograms in a reduced size viewing window. The device contains at least one point light source or line light source (which provides sufficiently coherent light), a lens, and a holographic array with cells arranged in a matrix with at least one opening per cell, the phase, or amplitude of the opening being controllable. A viewing plane is located in the image plane of the light source.
The hologram information is sampled in pixels and displayed on an LCD array. Sampled holograms always have the property of periodic repetitions of the reconstructed scene and the viewing window. Care has to be taken that the viewing windows do not overlap, as in that case multiple reconstructions would be seen. Limiting the area on the hologram on which the scene information is encoded can avoid an overlap. This area has to be limited such that light emanating from reconstructed scene points is confined to one viewing window. Therefore, the device reconstructs a video hologram in one periodicity interval of the Fourier transform in a viewing plane. The reconstructed three-dimensional scene can be viewed with both eyes through a viewing window located in front of the eyes. The reconstructed scene is visible inside a reconstruction frustum, which stretches between the display area and the viewing window; the scene can thereby be reconstructed on, in front of or behind the array surface. The known solution allows the use of a conventional array with resolution near 3 million pixels at reasonable hardware expenses and computing power.
A light source according to this document is considered sufficiently coherent if the light is spatially coherent to an extent that it allows interference, so that it is at least suitable for a one-dimensional holographic reconstruction with an adequate resolution. These requirements can also be met by conventional light sources, like an LED arrangement, if they radiate light through an adequately narrow gap. The spectral bandwidth of high-brightness LEDs is sufficient to ensure temporal coherence for holographic reconstruction. A line light source can be considered a point light source if seen from a right angle to its length. The light is then coherent in this direction and incoherent in the perpendicular direction. In order to ensure temporal coherence, the light must have an adequately narrow wavelength range. Color holograms can be displayed when the information may be divided spatially into spectral portions monochromatically, sequentially or by way of filter means. The electronically controllable pixels arranged in the holographic array can be an intensity-modulating SLM, a phase-modulating SLM or an SLM that modulates both the amplitude and the phase of the light capable of generating interference. Pixel arrays, which are unable to directly control the phase of the coherent light, like a conventional LCD, may use the known detour phase coding method so that amplitude settings with several controllable pixels per holographic image point control the light phase. For encoding a complex value for a single holographic image point of the array such known encoding technique uses three electronically controllable pixels.
In contrast to common known solutions, the present invention and the solution according to application WO 2004/044659 (A2) encodes the hologram information of a single scene point to a restricted encoded area of the holographic array only. The extension and position of the encoded area are chosen such that the reconstructed scene point is visible only within the viewing window. The observer cannot see the periodic repetitions of the reconstructed scene point, as the light emanating from those points does not reach the central viewing window. The extension and position of the encoded area depend on the x, y, and z coordinates of the scene points.
The pupil of each eye has to be located in the viewing window. Due to the smallness of the window, an eye position-tracking device detects the observer's eyes and controls the position of the viewing window according to the observer's movement. Vertical tracking is achieved by vertical shifting of the light source. This will shift the viewing window containing the reconstructed scene.
Another way to reduce the expense of reconstruction video holograms is to assign two separate viewing windows, each to one eye of the observer, achieved by two separate, adequately coherent light sources being alternately turned on and two separate holograms encoded synchronously with the switch-over of the light sources. The SLM alternately encodes the two video holograms displaying different perspective views. Due to the low refresh frequency and long switching delays of available hardware, a sequential representation leads however to cross talk between the two eyes.